LoLA is a LoL (League of Legends) game data analysis / analytics project.
The data crawling part is based on Riot API and a Python wrapper Cassiopeia. A SQLite database is designed and used in this project, which remodels and stores game objects for our analysis objectives. The database I/O part involves sqlite3 and pandas.
This part has been well tested with Python 3.5, though in some environments (e.g. Windows cmd) a decode/encode error may occur in print functions due to multi-language issue; you can just comment out all print codes without any influcence on crawling itself. Python 2.X may also run well with a few edits. In addition, since both Riot API and Cassiopeia are not quite stable for now, we may also update constantly to keep compatible.
We have obtained data of over 220,000 Ranked-SOLO-5x5 matches with details in the North American region, Pre-Season 2016. And we plan to swallow more in the future.
We are doing analyses such as:
- Champion Ability Rank
- Champion Clustering
- Champion Recommendation
- Winning Analysis and Match Prediction
- Cheating Detection
Interesting results will be uploaded continuously. As we are doing many experiments, code in this part is quite messy now and will be refined later.
We plan to share our database file later.
If you are interested in this project, feel free to participate in.